Archived posting to the Leica Users Group, 2007/03/14
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]A detailed article in LFI 2/2007 explains that the DNG file is not a true 8-bit file. Rather it is (my words) a logarithmically quantized 14-bit file, allowing significant 6:1 file compression (a 10 Mb DNG file opens to 59 Mb) It works like this (my Cliff Notes version): A true 8-bit file will quantize the luminance values into 2^8 = 256 equally spaced values. The key takeaway is that the steps are equally spaced, shadow to highlight. The DNG file starts with 2^14 = 16,384 steps, and then chooses 256 values, logarithmically spaced so that the shadows get most of the steps, and the highlights less. This lets the DNG use only 8-bits to encode the file. The steps can be interpolated back to 14 bits with the Adobe DNG converter. "It interprets the 8-bit data as an index to a table containing the actual tonal values in 14-bit resolution." Confused? What it all boils down to is that the M8's DNG files should get better tonal separation in the shadow areas than other camera RAW formats that use 12-bits, with acceptable separation on the highlights. Cheers, Bob Rose